Effects of blood donation on cerebral blood flow velocity C. SILVANI, A. ZANELLA, F. Rossi, G. CABRINI, C. FERRARI, B. CESANA,AND G. SIRCHIA Eighteen healthy male blood donors, nine with hematocrit (Hct) of 0.40 to 0.45 (normal Hct) and nine with Hct of 0.49 to 0.52 (upper-limit Hct), were monitored by continuous-wave internal carotid Doppler sonography and hematologic tests for 28 days after blood donation, to ascertain whether and to what extent a single standard donation may modify the velocity of cerebral blood flow. The two groups had similar mean predonation values of internal carotid flow velocity (ICFV): blood donation was followed in both groups by a slight, transient decrease of ICFV at the end of phlebotomy, due to donation-induced hypovolemia, and then by an increase of ICFV lasting 7 to 10 days. Analysis of individual profiles revealed that only four of nine upper-limit and six of nine normal Hct donors displayed a positive trend (increase) in the ICFV within the first week after donation, and that it was due mainly to a rise in systolic flow velocity. Mean Hct and arterial oxygen content showed a negative trend (decrease) within the first week that was opposite to the ICFV trend. Other laboratory variables, including serum proteins and plasma fibrin0 en concentration, and the iron status indicators did not change, except for serum !erritin, which also decreased within the first week after phlebotomy. It can be concluded that blood donation may result in a short-term increase of blood flow velocity that is independent of Hct predonation levels in approximately one-half of the donors. TRANSFUSION 1990;30:710-713.
IT IS KNOWN THAT S U B J E ~ Swith hematocrit (Hct) values over 0.48 are at increased risk of cerebrovascular Thomas et aL4 demonstrated that patients with high or high-normal Hct, whatever the cause, display a reduction of cerebral blood flow (CBF) that is probably one of the factors responsible for their increased incidence of stroke. Lowering Hct by repeated phlebotomy has been shown to improve CBF in polycythemic or upper-limit Hct patients,'n6 although it is not clear whether this effect is due mainly to the decrease of viscosity" or rather to the reduction of arterial oxygen content' that follows phlebotomy. At present, no information is available on the effect of standard blood donation on CBF in healthy subjects. The aim of the present study w a s to ascertain whether the donation of a single unit of blood may modify CBF, particularly in upper-limit Hct subjects who represent some 8 percent of repeat donors.8 For this purpose, w e monitored nine normal and nine upper-limit Hct blood donors at intervals before and after blood donation by evaluating the internal carotid flow velocity (ICFV), which is known to be related to CBF.' Hematologic variables related to blood viscosity and arterial oxygen content were also evaluated.
Materials and Methods Subjects We studied nine subjects with venous Hct of 0.49 to 0.52 (upper-limit Hct) confirmed on at least three different occasions and nine subjects with Hct of 0.40 to 0.45 (normal Hct). The two groups of subjects were matched for age (range, 2035 years; median, 25), body mass index (range, 20-26; median, 22), blood pressure, serum cholesterol, donation number (range, 3-20; median, 5), smoking, drinking, physical exercise, and life-style. All subjects were men who were accepted as blood donors on the basis of negative clinical history and normal results at both physical examination and routine laboratoiy tests, including serum glucose and cholesterol concentrations, platelet count, and red cell sedimentation rate. Upperlimit Hct blood donors had previously undergone additional tests (blood gas analysis, respiratory function, red cell mass and plasma volume measurements, screening for unstable and abnormal hemoglobins, and white cell alkaline phosphatase levels), as described elsewhere,8 to exclude possible causes of polycythernia. All the selected subjects gave a standard (400 mL) blood donation. Two days and 1 day before venesection, at its cnd, and 1, 3, 7, 10, 14, 21, and 28 days after, the ICFV was recorded and the following hematologic variables determined: complete blood cell count, including platelet and reticulocyte numbers and serum protein and plasma fibrinogen concentrations; serum iron concentration and total iron-binding ciipacity; and serum ferritin and red cell zinc protoporphyrin levels.
Technical procedures
From thc Ccntro Trasfusionalc c di Immunologia dei Trapianti, Clinica Ncurochirurgica and Laboratorio Biostatistico, Ospcdalc Maggiorc Policlinico, Milan, Italy. Supportcd in par1 by grant No. 86020G15G from tlic Consiglio Nazionalc dcllc Ricerchc, Romc, Italy. Rcceivcd for publication July 19, 1989; rcvision rcccived January 19, 1990, and acccptcd April G, 1990.
We determined the hemoglobin (Hb) concentration, Hct, blood cell counts, and red cell indices with an automated, laserbased cell counter (ELT-800, Ortho Diagnostic Systems, Westwood, MA). We performed the reticulocyte count in duplicate according to Dacie and Lewis'O and determined the serum iron concentration and total iron-binding capacity by the
710
TRANSFUSION 159O-Vot. 30. No. 8
711
BLOOD DONATION AND CEREBRAL BLOOD FLOW
colorimetric method (Ferrochem Model 3050 analyzer, Environmental Sciences Associates, Bedford, MA). We measured serum ferritin by radioimmunoassay (Lisophase, Sclavo, Siena, Italy) and red cell zinc protoporphyrin by direct fluorometry using a hematofluorometer (Model 4000, Environmental Sciences Associates), as described elsewhere. I I Serum protein and plasma fibrinogen concentrations were determined by the standard biuret test and according to the method of Blomback,12 respectively, Arterial oxygen content was calculated from venous Hb concentration by means of the tables of Kelman and Nunn.” We performed measurements of ICFV according to Barnes et aI.,l4 using a 5.6-MHz, continuous-wave, bidirectional Doppler detector (Model 1010; Parks Electronic Laboratories, Beaverton, OR). Systolic (ICFV-s), diastolic (ICFVd), and mean ICFVs were calculated by Doppler equation.” To reduce technical variability, each subject lay in the supine position for 10 minutes to stabilize heart rate and arterial pressure; when the best signal was identified, we recorded it and marked the corresponding point on the neck for the next rneasurements, which were taken by the same operator at the same point on the neck. Under these experimental conditions, the technical variability in each subject was 5 percent.“
-
J,
01,
Statistical analysis We used analysis of variance” for a mixed factorial design with repeated measurements on the same subject and corrected the significant level of variation within subjects according to Geisser and Greenhouse. Comparison against predonation values was performed according to multiple, a priori orthogonal comparisons; in addition, we split “between-time” variance according to the orthogonal polynomial method to test the presence of functional trends.17 For each subject, we defined a trend as at least three consecutive values progressively increasing (positive trend) or decreasing (negative trend) by at least 10 percent of basal value. Statistical analysis was performed with a software program (BMDP Statistical Software, h s Angeles; CA)” implemented on a computer (Model llOO/ 90, Univac, Roseville, MN).
Results Figure 1 shows the pattern of ICFV in upper-limit and normal Hct subjects before and after blood donation. The two groups had similar mean ICFV predonation values, and both displayed a slight, transient decrease in ICFV at the end of phlebotomy (p = 0.0007), followed by an increase (p LL
0 I
-4 I ,
.- 1
7 'r
1 c
0
x v ?
.20
>
4
LL
u I
-40
-4 I
)
0 1
I
i
3
7
10
14
21
28
0 1
3
7
1'0
14
2'1
28
1
TIME ( days FIG.2. Longitudinal individual profiles of systolic (fop) and diastolic (bottom) intcrnal carotid flow velocity (ICFV-s and ICFV-d, rcspectivcly) in ninc upper-limit hematocrit (Hct) (left) and ninc normal Hct (right) blood donors at various intervals after blood drawing. Timc 0 corrcsponds to thc cnd of blood donation. Flow vclocity is cxprcsscd as pcrcent of mcan prcdonation values.
finding that differs from the observations of Thomas et al.4 and Humphrey et al.,z' who measured blood flow by isotopic techniques and found higher values in normal Hct patients than in those with upper-limit Hct. Moreover, the ICFV values observed in our study are, on average, higher than those reported by Lindegaard et al.9 in neurologic patients with comparable Hct levels. After blood donation, ICFV showed a similar pattern in the two groups, both of which displayed a transient decrease at the end of phlebotomy, followed by an increase lasting 7 to 10 days. The slight reduction of ICFV immediately after donation, observed in almost all subjects, was probably due to hypovolemia induced by phlebotomy. The subsequent increase in mean ICFV seemed to be related in both groups to changes in Hct and arterial oxygen content, as was observed by Thomas4 and
Humphreyz6 and their coworkers in polycythemic patients and by Korosue et al.z7 in patients undergoing isovolemic hemodilution. When individual profiles are analyzed, it is interesting that only four of nine upperlimit and six of nine normal Hct donors displayed a positive trend in ICFV, and it was due mainly to a rise in ICFV-s. The reason for a significant postdonation rise in ICFV in some subjects, independent of the basal Hct values, is still unclear. The Hct profile was similar in the two groups, decreasing by as much as 10 percent of the basal value within 3 days and returning to predonation levels within 4 weeks, as was observed also by LiedCnz8 in first-time blood donor candidates and Challoner et aLz9 in a few polycythemic patients. Changes in the other variables examined, including fibrinogen concentration, were minimal or not significant (as also
TRANSFUSION 1990-Vol. 30, No. 8
BLOOD DONATION AND CEREBRAL BLOOD FLOW
reported by other^^^-^'), except for those in serum ferritin (SF). It is worth noting that, in both groups, SF showed an unexpectedly high decrease (as much as 50% of the predonation value) within 7 days after donation, with the higher decreases being observed in subjects with the higher basal SF concentrations. To our knowledge, the shortterm effects on SF concentration of a single blood donation have never been evaluated. The finding that SF decreased more than would be expected as a result of the iron loss due to phlebotoniy supports the hypothesis that SF concentration is related not only to iron stores but also to the labile iron In conclusion, blood donation results in an increase in ICFV in one-half of the donors, independent of predonation Hct levels. Further investigations are needed to ascertain whether these observations may have any practical implication for individuals at risk for cerebrovascular accidents. References 1. Kanncl WB, Gordon T, Wolf PA, McNamara P. Hcmoglobin and the risk of ccrebral infarction: the Framingham Study. Strokc 1972;3:409-20. 2. Tohgi H, Yamanouchi H, Murakami M, Kamcyama M. Importance of the hcmatocrit as a risk factor in ccrcbral infarction. Strokc 1978;9:369-74. 3. Lowc GD, Forbcs CD. Blood rhcology and thrombosis. Clin Hacmatol 1981;10:343-67. 4. Thomas DJ, Marshall J, Russcll RW, ct al. Effect of hacmatocrit on ccrebral blood-flow in man. Lancet 1977;2:941-3. 5 . Wade JP, Pearson TC, Russel RW, Wetherlcy-Mcin G. Cerebral blood flow and blood viscosity in patients with polycythacmia secondary to hypoxic lung diseasc. Br Mcd J [Clin Res] 1981;283:689-92. 6. Thomas DJ, du Boulay GH, Marshall J, ct al. Ccrcbrdl bloodflow in polycythacmia. Lincct 1977;2:161-3. 7. Brown MM, Miirshall J. Rcgulation of ccrcbral blood flow in rcsponsc to changcs in blood viscosity. Lancct 1985;1:604-9. 8. Zanclla A, Silvani C, Banfi P, Bcllonc A, Fumagalli G, Sirchia G. Scrccning and evaluation of blood donors with uppcr-limit hcmatocrit Icvels. Transfusion 1987;27:485-7. 9. Lindcgaard KF, Lundar T, Wibcrg J, Sjobcrg D, Aaslid R, Norncs H. Variations in middlc ccrcbral artcry blood flow invcstigatcd with noninvasivc transcranial blood velocity measurements. Strokc 1987;18:1025-30. 10. Dacic JV, Lcwis SM. Practical hacmatology. 5th cd London: Churchill, 1975. 11. Zanclla A, Gridclli L, Bcrzuini A, et al. Sensitivity and prcdictive valuc of scrum ferritin and free erythrocyte protoporphyrin for iron dcficicncy. J Lab Clin Mcd 1989;113:73-8. 12. Blomback B. On the propcrtics of fibrinogcn and fibrin. Arkiv Kcmi 1958;12:99-113. 13. Kelman GR, Nunn JF. Computer produccd physiological tables for calculations involving the relationships betwecn blood oxygen tension and content. London: Buttenvorths, 1968. 14. Barnes RW, Rittgcrs SE, Putncy WW. Rcal-timc Dopplcr spcctrum analysis: prcdictivc valuc in defining operable carotid artery diseasc. Arch Surg 1982;117:52-7.
713
15. Blackshear WM, Phillips DJ, Chikos PM, Harlcy JD, Thick BL, Strandncss DE. Carotid artcry velocity pattcrns in normal and stcnotic vcsscls. Stroke 1980;11:67-71. 16. Cabrini GP, Fcrrari C, Rocca A, Ccsana B. [Rcproducibility of Dopplcr signals in the intcrnal carotid. Prcliminary study.] Clinic 1984;2:105-8. 17. Wincr BJ. Statistical principles in cxpcrirncntal dcsign. 2nd cd. New York: McGraw-Hill, 1971. 18. Gcisscr S , Grcenhouse SW. On mcthods in tlic analysis of profilc data. Psychometrica 1959;24:95-112. 19. Dixon WJ, cd. BMDP statistical softwarc: 1981. Bcrkclcy: University of California Press, 1981. 20. Humphrey PRD, Wade JPH. Oxygen transport, blood viscosily and ccrcbral blood flow. In: Schmid-Schonbcin H, Mcssmcr K, Ricgcr H, eds. Hcmodilution and flow improvcmenf. Bascl: Kargcr, 1981:145-8. 21. Wade JP. Transport of oxygen to thc brain in paficnts with clcvated hacmatocrit values before and aftcr vcncscction. Brain 1983;106:513-23. 22. Humphrey PRD, Michael J, Pcarson TC. Managcmcnt of rclativc polycythaemia: studies of cerebral blood flow and viscosity. Br J Haematol 1980;46:427-33. 23. Humphrey PRD, du Boulay GH, Marshall J, ct al. Ccrcbral bloodflow and viscosity in relative polycythaemia. Lanccf 1979;2:873-7. 24. Lieden G. Iron state in rcgular blood donors. Scand J Hacm;iiol 1973;11:342-9. 25. Wardsworth GR. Recovely from acute hacmorhage in normal mcn and women. J Physiol (London) 1955;129:583. 26. Humphrey PR. Changes in cerebral blood flow rclating to hacmatocrit and Viscosity. S a n d J Clin Lab Invest Suppl 1961;41:20911. 27. Korosue K, lshida K, Matsuoka H, Nagao T, Tamaki N, Matsumoto S . Clinical, hcmodynamic, and hcnicirhcological cffccts of isovolcmic hernodilution in acutc ccrcbral infarction. Ncurosurgcry 1988;23:148-53. 28. LiedCn G, Hoglund S, Ehn L. Changcs in certain iron mctabolism variablcs after a single blood donation. Acfa Mcd Scand 1975;197:27-30. 29. Challoner T, Briggs C, Rampling MW, Thomas DJ. A siudy of the hacmatological and haemorheological conscqucnccs of vciicsection. Br J Haematol 1986;62:671-8. 30. Hillman RS. Erythrocyfe disorders-ancniias duc to acutc blood loss. In Williams WJ, Bcutlcr E, Erslcv AJ, Lichtman MA, cds. Hcmatology. 3rd cd. Ncw‘York: McGraw-Hill, 3985:667-71. 31. Millcr ME, Cronkite EP, Garcia JF. Plasma lcvcls of ininiunoreactivc crythropoictin after acute blood loss in m i i n . Br J Hiicmatol 1982;52:545-9. 32. Jacobs A. LOW molecular rate intraccllular iron transport compounds. Blood 1977;50:433-9. Carla Silvani, MD, Scnior Assistant, Ccntro Trasfusionalc c di Inimunologia dci Trapianti, Ospcdalc Policlinico, via Franccsco Sforza 35, 20122 Milan, Italy. [Reprint request] Albcrto Zanclla, MD, Senior Assistant, Ccntro Trasfusionalc c di Immunologia dci Trapianti. Fabio Rossi, MD, Assistant, Centro Trasfusionalc c di 1mmunologi;i dei Trapianti. Gianpicto Cabrini, MD, Senior Assistant, Clinicii Ncurochirurgica dcll’ UniversitB, Ospedale Policlinico. Camillo Fcrrari, MD, Assistant, Clinica Ncurochirurgica dcll’ UniversitP. Bruno Ccsana, MD, Assistant, Laboratorio Biostatisfico, Ospcdalc Policlinico. Girolamo Sirchia, MD, Director, Centro Trasfusionalc c di Immunologia dei Trapianti.
IT IS KNOWN THAT S U B J E ~ Swith hematocrit (Hct) values over 0.48 are at increased risk of cerebrovascular Thomas et aL4 demonstrated that patients with high or high-normal Hct, whatever the cause, display a reduction of cerebral blood flow (CBF) that is probably one of the factors responsible for their increased incidence of stroke. Lowering Hct by repeated phlebotomy has been shown to improve CBF in polycythemic or upper-limit Hct patients,'n6 although it is not clear whether this effect is due mainly to the decrease of viscosity" or rather to the reduction of arterial oxygen content' that follows phlebotomy. At present, no information is available on the effect of standard blood donation on CBF in healthy subjects. The aim of the present study w a s to ascertain whether the donation of a single unit of blood may modify CBF, particularly in upper-limit Hct subjects who represent some 8 percent of repeat donors.8 For this purpose, w e monitored nine normal and nine upper-limit Hct blood donors at intervals before and after blood donation by evaluating the internal carotid flow velocity (ICFV), which is known to be related to CBF.' Hematologic variables related to blood viscosity and arterial oxygen content were also evaluated.
Materials and Methods Subjects We studied nine subjects with venous Hct of 0.49 to 0.52 (upper-limit Hct) confirmed on at least three different occasions and nine subjects with Hct of 0.40 to 0.45 (normal Hct). The two groups of subjects were matched for age (range, 2035 years; median, 25), body mass index (range, 20-26; median, 22), blood pressure, serum cholesterol, donation number (range, 3-20; median, 5), smoking, drinking, physical exercise, and life-style. All subjects were men who were accepted as blood donors on the basis of negative clinical history and normal results at both physical examination and routine laboratoiy tests, including serum glucose and cholesterol concentrations, platelet count, and red cell sedimentation rate. Upperlimit Hct blood donors had previously undergone additional tests (blood gas analysis, respiratory function, red cell mass and plasma volume measurements, screening for unstable and abnormal hemoglobins, and white cell alkaline phosphatase levels), as described elsewhere,8 to exclude possible causes of polycythernia. All the selected subjects gave a standard (400 mL) blood donation. Two days and 1 day before venesection, at its cnd, and 1, 3, 7, 10, 14, 21, and 28 days after, the ICFV was recorded and the following hematologic variables determined: complete blood cell count, including platelet and reticulocyte numbers and serum protein and plasma fibrinogen concentrations; serum iron concentration and total iron-binding ciipacity; and serum ferritin and red cell zinc protoporphyrin levels.
Technical procedures
From thc Ccntro Trasfusionalc c di Immunologia dei Trapianti, Clinica Ncurochirurgica and Laboratorio Biostatistico, Ospcdalc Maggiorc Policlinico, Milan, Italy. Supportcd in par1 by grant No. 86020G15G from tlic Consiglio Nazionalc dcllc Ricerchc, Romc, Italy. Rcceivcd for publication July 19, 1989; rcvision rcccived January 19, 1990, and acccptcd April G, 1990.
We determined the hemoglobin (Hb) concentration, Hct, blood cell counts, and red cell indices with an automated, laserbased cell counter (ELT-800, Ortho Diagnostic Systems, Westwood, MA). We performed the reticulocyte count in duplicate according to Dacie and Lewis'O and determined the serum iron concentration and total iron-binding capacity by the
710
TRANSFUSION 159O-Vot. 30. No. 8
711
BLOOD DONATION AND CEREBRAL BLOOD FLOW
colorimetric method (Ferrochem Model 3050 analyzer, Environmental Sciences Associates, Bedford, MA). We measured serum ferritin by radioimmunoassay (Lisophase, Sclavo, Siena, Italy) and red cell zinc protoporphyrin by direct fluorometry using a hematofluorometer (Model 4000, Environmental Sciences Associates), as described elsewhere. I I Serum protein and plasma fibrinogen concentrations were determined by the standard biuret test and according to the method of Blomback,12 respectively, Arterial oxygen content was calculated from venous Hb concentration by means of the tables of Kelman and Nunn.” We performed measurements of ICFV according to Barnes et aI.,l4 using a 5.6-MHz, continuous-wave, bidirectional Doppler detector (Model 1010; Parks Electronic Laboratories, Beaverton, OR). Systolic (ICFV-s), diastolic (ICFVd), and mean ICFVs were calculated by Doppler equation.” To reduce technical variability, each subject lay in the supine position for 10 minutes to stabilize heart rate and arterial pressure; when the best signal was identified, we recorded it and marked the corresponding point on the neck for the next rneasurements, which were taken by the same operator at the same point on the neck. Under these experimental conditions, the technical variability in each subject was 5 percent.“
-
J,
01,
Statistical analysis We used analysis of variance” for a mixed factorial design with repeated measurements on the same subject and corrected the significant level of variation within subjects according to Geisser and Greenhouse. Comparison against predonation values was performed according to multiple, a priori orthogonal comparisons; in addition, we split “between-time” variance according to the orthogonal polynomial method to test the presence of functional trends.17 For each subject, we defined a trend as at least three consecutive values progressively increasing (positive trend) or decreasing (negative trend) by at least 10 percent of basal value. Statistical analysis was performed with a software program (BMDP Statistical Software, h s Angeles; CA)” implemented on a computer (Model llOO/ 90, Univac, Roseville, MN).
Results Figure 1 shows the pattern of ICFV in upper-limit and normal Hct subjects before and after blood donation. The two groups had similar mean ICFV predonation values, and both displayed a slight, transient decrease in ICFV at the end of phlebotomy (p = 0.0007), followed by an increase (p LL
0 I
-4 I ,
.- 1
7 'r
1 c
0
x v ?
.20
>
4
LL
u I
-40
-4 I
)
0 1
I
i
3
7
10
14
21
28
0 1
3
7
1'0
14
2'1
28
1
TIME ( days FIG.2. Longitudinal individual profiles of systolic (fop) and diastolic (bottom) intcrnal carotid flow velocity (ICFV-s and ICFV-d, rcspectivcly) in ninc upper-limit hematocrit (Hct) (left) and ninc normal Hct (right) blood donors at various intervals after blood drawing. Timc 0 corrcsponds to thc cnd of blood donation. Flow vclocity is cxprcsscd as pcrcent of mcan prcdonation values.
finding that differs from the observations of Thomas et al.4 and Humphrey et al.,z' who measured blood flow by isotopic techniques and found higher values in normal Hct patients than in those with upper-limit Hct. Moreover, the ICFV values observed in our study are, on average, higher than those reported by Lindegaard et al.9 in neurologic patients with comparable Hct levels. After blood donation, ICFV showed a similar pattern in the two groups, both of which displayed a transient decrease at the end of phlebotomy, followed by an increase lasting 7 to 10 days. The slight reduction of ICFV immediately after donation, observed in almost all subjects, was probably due to hypovolemia induced by phlebotomy. The subsequent increase in mean ICFV seemed to be related in both groups to changes in Hct and arterial oxygen content, as was observed by Thomas4 and
Humphreyz6 and their coworkers in polycythemic patients and by Korosue et al.z7 in patients undergoing isovolemic hemodilution. When individual profiles are analyzed, it is interesting that only four of nine upperlimit and six of nine normal Hct donors displayed a positive trend in ICFV, and it was due mainly to a rise in ICFV-s. The reason for a significant postdonation rise in ICFV in some subjects, independent of the basal Hct values, is still unclear. The Hct profile was similar in the two groups, decreasing by as much as 10 percent of the basal value within 3 days and returning to predonation levels within 4 weeks, as was observed also by LiedCnz8 in first-time blood donor candidates and Challoner et aLz9 in a few polycythemic patients. Changes in the other variables examined, including fibrinogen concentration, were minimal or not significant (as also
TRANSFUSION 1990-Vol. 30, No. 8
BLOOD DONATION AND CEREBRAL BLOOD FLOW
reported by other^^^-^'), except for those in serum ferritin (SF). It is worth noting that, in both groups, SF showed an unexpectedly high decrease (as much as 50% of the predonation value) within 7 days after donation, with the higher decreases being observed in subjects with the higher basal SF concentrations. To our knowledge, the shortterm effects on SF concentration of a single blood donation have never been evaluated. The finding that SF decreased more than would be expected as a result of the iron loss due to phlebotoniy supports the hypothesis that SF concentration is related not only to iron stores but also to the labile iron In conclusion, blood donation results in an increase in ICFV in one-half of the donors, independent of predonation Hct levels. Further investigations are needed to ascertain whether these observations may have any practical implication for individuals at risk for cerebrovascular accidents. References 1. Kanncl WB, Gordon T, Wolf PA, McNamara P. Hcmoglobin and the risk of ccrebral infarction: the Framingham Study. Strokc 1972;3:409-20. 2. Tohgi H, Yamanouchi H, Murakami M, Kamcyama M. Importance of the hcmatocrit as a risk factor in ccrcbral infarction. Strokc 1978;9:369-74. 3. Lowc GD, Forbcs CD. Blood rhcology and thrombosis. Clin Hacmatol 1981;10:343-67. 4. Thomas DJ, Marshall J, Russcll RW, ct al. Effect of hacmatocrit on ccrebral blood-flow in man. Lancet 1977;2:941-3. 5 . Wade JP, Pearson TC, Russel RW, Wetherlcy-Mcin G. Cerebral blood flow and blood viscosity in patients with polycythacmia secondary to hypoxic lung diseasc. Br Mcd J [Clin Res] 1981;283:689-92. 6. Thomas DJ, du Boulay GH, Marshall J, ct al. Ccrcbrdl bloodflow in polycythacmia. Lincct 1977;2:161-3. 7. Brown MM, Miirshall J. Rcgulation of ccrcbral blood flow in rcsponsc to changcs in blood viscosity. Lancct 1985;1:604-9. 8. Zanclla A, Silvani C, Banfi P, Bcllonc A, Fumagalli G, Sirchia G. Scrccning and evaluation of blood donors with uppcr-limit hcmatocrit Icvels. Transfusion 1987;27:485-7. 9. Lindcgaard KF, Lundar T, Wibcrg J, Sjobcrg D, Aaslid R, Norncs H. Variations in middlc ccrcbral artcry blood flow invcstigatcd with noninvasivc transcranial blood velocity measurements. Strokc 1987;18:1025-30. 10. Dacic JV, Lcwis SM. Practical hacmatology. 5th cd London: Churchill, 1975. 11. Zanclla A, Gridclli L, Bcrzuini A, et al. Sensitivity and prcdictive valuc of scrum ferritin and free erythrocyte protoporphyrin for iron dcficicncy. J Lab Clin Mcd 1989;113:73-8. 12. Blomback B. On the propcrtics of fibrinogcn and fibrin. Arkiv Kcmi 1958;12:99-113. 13. Kelman GR, Nunn JF. Computer produccd physiological tables for calculations involving the relationships betwecn blood oxygen tension and content. London: Buttenvorths, 1968. 14. Barnes RW, Rittgcrs SE, Putncy WW. Rcal-timc Dopplcr spcctrum analysis: prcdictivc valuc in defining operable carotid artery diseasc. Arch Surg 1982;117:52-7.
713
15. Blackshear WM, Phillips DJ, Chikos PM, Harlcy JD, Thick BL, Strandncss DE. Carotid artcry velocity pattcrns in normal and stcnotic vcsscls. Stroke 1980;11:67-71. 16. Cabrini GP, Fcrrari C, Rocca A, Ccsana B. [Rcproducibility of Dopplcr signals in the intcrnal carotid. Prcliminary study.] Clinic 1984;2:105-8. 17. Wincr BJ. Statistical principles in cxpcrirncntal dcsign. 2nd cd. New York: McGraw-Hill, 1971. 18. Gcisscr S , Grcenhouse SW. On mcthods in tlic analysis of profilc data. Psychometrica 1959;24:95-112. 19. Dixon WJ, cd. BMDP statistical softwarc: 1981. Bcrkclcy: University of California Press, 1981. 20. Humphrey PRD, Wade JPH. Oxygen transport, blood viscosily and ccrcbral blood flow. In: Schmid-Schonbcin H, Mcssmcr K, Ricgcr H, eds. Hcmodilution and flow improvcmenf. Bascl: Kargcr, 1981:145-8. 21. Wade JP. Transport of oxygen to thc brain in paficnts with clcvated hacmatocrit values before and aftcr vcncscction. Brain 1983;106:513-23. 22. Humphrey PRD, Michael J, Pcarson TC. Managcmcnt of rclativc polycythaemia: studies of cerebral blood flow and viscosity. Br J Haematol 1980;46:427-33. 23. Humphrey PRD, du Boulay GH, Marshall J, ct al. Ccrcbral bloodflow and viscosity in relative polycythaemia. Lanccf 1979;2:873-7. 24. Lieden G. Iron state in rcgular blood donors. Scand J Hacm;iiol 1973;11:342-9. 25. Wardsworth GR. Recovely from acute hacmorhage in normal mcn and women. J Physiol (London) 1955;129:583. 26. Humphrey PR. Changes in cerebral blood flow rclating to hacmatocrit and Viscosity. S a n d J Clin Lab Invest Suppl 1961;41:20911. 27. Korosue K, lshida K, Matsuoka H, Nagao T, Tamaki N, Matsumoto S . Clinical, hcmodynamic, and hcnicirhcological cffccts of isovolcmic hernodilution in acutc ccrcbral infarction. Ncurosurgcry 1988;23:148-53. 28. LiedCn G, Hoglund S, Ehn L. Changcs in certain iron mctabolism variablcs after a single blood donation. Acfa Mcd Scand 1975;197:27-30. 29. Challoner T, Briggs C, Rampling MW, Thomas DJ. A siudy of the hacmatological and haemorheological conscqucnccs of vciicsection. Br J Haematol 1986;62:671-8. 30. Hillman RS. Erythrocyfe disorders-ancniias duc to acutc blood loss. In Williams WJ, Bcutlcr E, Erslcv AJ, Lichtman MA, cds. Hcmatology. 3rd cd. Ncw‘York: McGraw-Hill, 3985:667-71. 31. Millcr ME, Cronkite EP, Garcia JF. Plasma lcvcls of ininiunoreactivc crythropoictin after acute blood loss in m i i n . Br J Hiicmatol 1982;52:545-9. 32. Jacobs A. LOW molecular rate intraccllular iron transport compounds. Blood 1977;50:433-9. Carla Silvani, MD, Scnior Assistant, Ccntro Trasfusionalc c di Inimunologia dci Trapianti, Ospcdalc Policlinico, via Franccsco Sforza 35, 20122 Milan, Italy. [Reprint request] Albcrto Zanclla, MD, Senior Assistant, Ccntro Trasfusionalc c di Immunologia dci Trapianti. Fabio Rossi, MD, Assistant, Centro Trasfusionalc c di 1mmunologi;i dei Trapianti. Gianpicto Cabrini, MD, Senior Assistant, Clinicii Ncurochirurgica dcll’ UniversitB, Ospedale Policlinico. Camillo Fcrrari, MD, Assistant, Clinica Ncurochirurgica dcll’ UniversitP. Bruno Ccsana, MD, Assistant, Laboratorio Biostatisfico, Ospcdalc Policlinico. Girolamo Sirchia, MD, Director, Centro Trasfusionalc c di Immunologia dei Trapianti.
